Universally consistent hybrid regression model for water quality prediction

Abstract

In this work, we propose a hybrid regression model to solve a specific problem faced by a modern paper manufacturing company. Boiler inlet water quality is a major concern for the company since it helps to produce power and steam for the paper machine. If water treatment plant can not produce water of desired quality as specified by the boiler, then it results in poor health of the boiler water tube and consequently affects the quality of the paper. Variation in inlet water quality of the boiler is due to several crucial process parameters. We build a hybrid regression model for boiler water quality prediction based on decision trees and artificial neural networks. This model can be useful for manufacturing process quality improvement for the paper company. We have proved the desired statistical consistency of the hybrid model to show its robustness and universal use. The primary advantage of the model is its natural interpretability and excellent performance when compared with other state-of-the-art.